Focus on approval, commercialization How many points should be passed in medical imaging AI?

category:Finance
 Focus on approval, commercialization How many points should be passed in medical imaging AI?


So far this year, only we and Shukun technology have obtained two rounds of financing in the industry. At the meeting, the staff of Keya medical told the reporter of the science and technology innovation board daily. According to relevant data, as of August this year, the financing amount related to the medical imaging AI field totaled 1.56 billion yuan, and the financing amount of Keya Medical Co., Ltd. + Shukun Medical Co., Ltd. was close to 1 / 2 of the total financing amount.

From fanaticism to rationality, what is the future path of medical imaging AI enterprises? How many hurdles to pass?

The first level: is there a card in the approval?

On January 15 this year, the State Food and Drug Administration issued the first medical AI class III Certificate in China, and Keya medicals deep pulse score based on deep neural network was employed in clinical application.

This is an important node. With the first node, there is a standard.

He Weigang, director of the second Department of the medical device evaluation center of the State Drug Administration, also said at the second China Medical Imaging AI conference: for the new generation of artificial intelligence medical devices, the regulatory authorities have established a unified consensus to speed up the development of a new generation of artificial intelligence technology, and provide reasonable rule definition for the development and construction of medical artificial intelligence devices.

He Weigang also said: at present, the food and drug administration has established a special medical AI audit team, and released a number of review points on AI medical devices, which have provided help for AI medical devices to be approved as soon as possible.

Since Keya medical, the FDA has successively issued four medical AI class III certificates this year. The products include the ECG analysis software of Lepu medical, the magnetic resonance imaging aided diagnosis software of Ender technology, and the two fundus image aided diagnostic software of silicon-based intelligent and Eagle pupil medical for diabetic retinopathy. Up to now, there are five three kinds of certificates in the AI medical device industry.

Therefore, he Weigang believes that the development of medical AI at this stage is not in the process of approval.

The second level: is it difficult to commercialize?

After the approval threshold is stepped down, the commercialization pressure of enterprises will only be greater and greater.

AI is still in the promotion stage, and it is basically used by the top three hospitals, an investor in the medical industry told the reporter of the science and technology innovation board

Liu Shiyuan disclosed at the medical imaging AI conference that at present, the penetration rate of medical AI deployment in tertiary hospitals is 4.5% - 7%, which may reach 15-20% in the future. The epidemic has also played an accelerating role. Liu Shiyuan takes the use of pulmonary nodule products in Changzheng Hospital as an example: the utilization rate of Changzheng Hospital in 2017 is about 60%, and that in 2019 is 70%. After the end of the epidemic this year, the click through rate has reached more than 80%, which is more and more used by doctors.

Whats more, how to charge for the products entering the hospital?

On this issue, the aforementioned medical industry investors said: one is that the charging standard is not clear, and there is no medical insurance included.

In addition, as mentioned above, there are two AI screening software for sugar net to obtain three types of syndromes. According to relevant media reports, sugar net screening has been included in the list of charges of some hospitals, and sugar net AI enterprises can share it from hospitals. For example, the price of monocular examination in Chongqing is 33 yuan, while that in Guangzhou is 40 yuan. The total number of single eye examinations per day can reach 5000 cases, of which the share proportion of enterprises is about 50%.

It is worth mentioning here that if public hospitals want to charge, they must enter the charging catalogue. Liu Shiyuan revealed at the medical imaging AI conference that some provinces have begun to set up charging catalogue for AI products. However, there is still a long way to go from the establishment to the implementation.

In the insurance sector, as it is difficult to enter the medical insurance system, some commercial insurance companies may join in first in the future. Previously, the reporter of kechuangboard daily learned at the exchange meeting of medi technology that Bohai property insurance will cooperate with Madi technology in the field of assisted reproductive AI.

Third, how should enterprises and hospitals bear legal responsibilities?

He Lianhong, a professor and doctoral supervisor of the Law School of Central South University, said in politics and law that according to the causes of AI diagnostic errors in medical imaging, large hospitals, as owners and users of medical imaging detection equipment, may become the subject of liability, and medical imaging software designers, equipment manufacturers, equipment sellers, third parties or other civil owners The body may also become the subject of responsibility. There may be many responsible subjects involved in a diagnosis error. How to identify it? Is there any industry standard? These are the problems that need to be discussed at present. Since the development of medical imaging AI industry, it has experienced the craziest period of capital. The wind has stopped, and the challenges such as approval, commercialization and security issues are constantly on the rise However, it is because of the world is not determined that all the Heroes rise together. At the medical imaging AI conference, reporters from kechuangboard daily found that GPS (GE, Philips and Siemens), the three giants in the medical equipment industry, still occupy the most central position in the exhibition area. In China, there are Lianying, Neusoft medical, Shukun technology, Keya medical, Ender medical, ETO medical, Tui technology and other AI based start-up companies. Source: Science and technology innovation board daily editor: Zhang Mei_ NF2100

There may be many responsible subjects involved in a diagnosis error. How to identify it? Is there any industry standard? These are the problems that need to be discussed at present.

Since the development of medical imaging AI industry, it has experienced the craziest period of capital. The wind has stopped, and the challenges such as approval, commercialization and security issues are constantly on the rise

However, it is because of the world is not determined that all the Heroes rise together. At the medical imaging AI conference, reporters from kechuangboard daily found that GPS (GE, Philips and Siemens), the three giants in the medical equipment industry, still occupy the most central position in the exhibition area. In China, there are Lianying, Neusoft medical, Shukun technology, Keya medical, Ender medical, ETO medical, Tui technology and other AI based start-up companies.